The rise of commodity index traders (CITs) in the early 2000s provides a natural experiment to identify whether passive holding of long agricultural futures positions earns a positive risk premium. We use nearly a decade of daily nonpublic position data for all large traders to compute trading profits in twelve agricultural futures markets. Despite increasing price trends in a majority of markets, CITs were the biggest losers during the sample period, experiencing losses in nine out of twelve markets and an aggregate loss of $6.9 billion. This is just the opposite of the prediction of the theory of normal backwardation.